Immune systems of living organisms have an elaborate surveillance function to distinguish normal cells and abnormal cells in the body of themselves, and remove only the abnormal cells. However, when the surveillance function collapses, abnormal cells produced by mutation and the like cannot be removed but proliferate in the body. A mass of such proliferated abnormal cells is a tumor, i.e., cancer.
Cancer is mainly treated by surgical removal of cancer or use of anti-cancer agents. However, these treatment methods place physical burden due to removal surgery or side effects of anti-cancer agents, as well as mental burden due to operative scar.
In such background, a treatment method using an immunotherapy in combination is drawing attention. In the immunotherapy, cancer cells are attacked by increasing the number of immunocytes in patients themselves, and activating them. If the size of tumor formed by cancer cells can be reduced, the physical burden due to the removal surgery becomes small. In addition, since the operative scar is small, the mental burden is drastically reduced.
Natural killer (NK) T cells are immunocytes belonging to a novel lymphocyte lineage showing characteristics different from those of other lymphocyte lineages (T, B, NK cells). Since cytotoxic perforin granules are present in NKT cells, they are analogous to NK cells (non-patent document 1). However, since NKT cells express not only NK cell marker but also T cell receptor (TCR), it is clear that they form a definitively different, new cell group (non-patent document 2). NKT cells can produce both Th-1 type cytokine (mainly interferon (IFN)-γ) produced by helper T (Th)-1 cell that promotes immunostimulatory action and Th-2 type cytokine (mainly interleukin (IL)-4) produced by Th-2 cell that promotes immunosuppressive action (non-patent document 3), which suggests a possibility of controlling the balance of immune system (non-patent document 4). Therefore, by controlling the function of NKT cells, disrupted balance of the immune system is controlled and the surveillance function is enhanced, whereby cancer can be treated.
The most noticeable characteristic of NKT cells is that the α chain of TCR expressed by NKT cells is common to all members of one species. In other words, this means that all NKT cells of the living organisms belonging to the same species are activated by the same substance. This α chain is Vα24 in human and Vα14 in mouse, and they show extremely high homology between the two species. In addition, only very limited kinds of β chain are known to form a pair with the α chain. For this reason, this TCR is also called a “non-variable TCR”.
There are various kinds of glycosphingolipids which are known to be present in the body. In glycosphingolipids in the body, various sugars generally form a β-bond with ceramide. While the existent amount thereof varies depending on the organ, they are present in the cellular membrane of various organs (non-patent document 5).
In the meantime, a report has recently been documented that glycosphingolipids wherein sugar forms an α-bond with ceramide has a strong immunostimulatory action and an antitumor activity. α-Galactosylceramide represented by Agelasphins is a glycolipid isolated from an extract of Agelas mauritianus, one kind of sponge, and is known to strongly activate NKT cells (non-patent document 6).
After intake by antigen presenting cell (APC), which is represented by dendritic cell (DC) and the like, α-galactosylceramide is presented on the cellular membrane by a CD1d protein similar to major histocompatible complex (MHC) class I molecule. NKT cells are activated by recognition using TCR of the thus-presented complex of CD1d protein and α-galactosylceramide, which triggers various immune reactions.
α-Galactosylceramide is glycosphingolipids wherein galactose is bonded by α-configuration to a ceramide formed by acylation of sphingosine base with long chain fatty acid. Various analogs have been synthesized heretofore, and the correlation between structures and activities thereof has been investigated. It has been clarified that, in a series of synthesis analogs, for example, α-galactosylceramide represented by the following formula (a) (hereinafter to be referred to as “α-GalCer”) shows the strongest activity, and further, that the corresponding β-configuration (β-GalCer) does not show an immunostimulatory activity (non-patent document 7).

In recent years, therapeutic drugs containing α-GalCer as an active ingredient have been proposed or developed, taking note of such function of NKT cells. However, NKT cells activated by the administration of α-GalCer simultaneously produce, along with the production of IFN-γ, which is a cytokine that induces an immunostimulatory activity and is useful for cancer treatment, IL-4, which is a cytokine that induces an immunosuppressive action, and IL-10, which is a cytokine that induces an immunity regulating action. As a result, problems occur since immunostimulatory activity is suppressed and a sufficient effect for cancer treatment is difficult to provide.
In recent years, a glycolipid (α-C-GalCer) that preferentially produces IFN-γ, which is a cytokine that induces an immunostimulatory action of NKT cell, has been developed (patent documents 1-3, non-patent document 8). α-C-GalCer is an analog wherein the oxygen atom forming a glucoside bond of α-GalCer is substituted by a methylene group. It has been reported that the in vivo stability is enhanced and the efficacy is maintained for a long time since, in α-C-GalCer, the bond between sugar and ceramide is converted from a glycoside bond to a carbon-carbon bond (non-patent document 9). However, the reason for preferential production of IFN-γ by α-C-GalCer has not been elucidated yet. In addition, the IFN-γ/IL-4 ratio is not entirely sufficient from the clinical aspects and further improvement is desired.    patent document 1: US-A-2005/0222048    patent document 2: WO2003/105769    patent document 3: DE-A-10128250    non-patent document 1: Proc. Natl. Acad. Sci. USA 1998, 95, 5690-5693    non-patent document 2: J. Immunol. 1995, 155, 2972-2983    non-patent document 3: J. Immunol. 1998, 161, 3271-3281    non-patent document 4: Nat. Immunol. 2003, 4, 1164-1165    non-patent document 5: Biochim. Biophys. Acta 1973, 315-335    non-patent document 6: Science 1997, 278, 1626-1629    non-patent document 7: J. Med. Chem. 1995, 38, 2176-2187    non-patent document 8: Angew. Chem. Int. Ed. Engl. 2004, 43, 3818-3822    non-patent document 9: J. Exp. Med. 2003, 198, 1631-1641